Jul 5, 2026 - eRacks Restructures Private AI Server Line Into a Good, Better, Best Ladder, 4U Models From $16,995
Fremont, CA - Jul 5, 2026
FOR IMMEDIATE RELEASE
eRacks Restructures Private AI Server Line Into a Good, Better, Best Ladder, 4U Models From $16,995
New Good, Better, Best ladder pairs current-generation AMD Threadripper and EPYC processors with a pre-installed open-source AI stack, and cuts mid-tier prices substantially from the prior line.
FREMONT, CA, July 5, 2026: eRacks Open Source Systems, a builder of open-source rackmount servers since 1999, today relaunched its private AI server line as a Good, Better, Best ladder of eRacks-Certified workhorse rackmount servers, backed by an enterprise engineered-OEM tier. Every system runs large language models (LLMs, the software behind AI assistants) entirely on hardware the customer owns. Data never leaves the building, and there are no per-seat or per-token fees.
The Workhorse tier is fully enclosed rackmount servers, assembled, burned in, and certified by eRacks. The AILSA 2U (a U is 1.75 inches of rack height) starts at $5,995, uses Intel Arc GPUs (graphics processing units, the chips that run AI models), and handles 70B-class models (roughly 70 billion parameters). The new AISLING 4U starts at $16,995 with a 24-core AMD Threadripper 9960X, 128GB of ECC memory (error-correcting code memory, which detects and fixes memory errors), a single 1600W power supply, and up to three dual-slot GPUs for 96GB of VRAM (the GPU's onboard memory, which holds the model) using Intel Arc Pro B70 cards.
The new AILEEN 4U starts at $21,995 and moves to a server-class 32-core AMD EPYC 9355 with 12 memory channels, 192GB of ECC memory, up to four GPUs for 128GB of VRAM, and redundant 1+1 power (two supplies, either one can run the system). AILEEN is also available in custom colors: blue, black, white, or red.
For buyers who want validated OEM server systems with out-of-band management (remote hardware control that works even when the operating system is down), the enterprise tier offers the AIDAN 2U EPYC from $13,895 and the AISHA 4U from $30,995, with support for up to 10 GPUs.
Every system ships with Ubuntu LTS and the open-source AI stack pre-installed: Ollama, Open WebUI, vLLM, llama.cpp, and PyTorch. Each unit is burned in (run under sustained load before shipping) and tested, is air-gap ready (able to operate with no internet connection), and gives users browser access to their own AI models on day one.
"Buyers kept asking the same question: which box runs my model," said Joseph Wolff, Founder and CTO of eRacks Systems. "The new ladder answers it. Pick your model size, pick your tier, and the system arrives ready to use." Mid-tier prices dropped substantially from the prior line, reflecting current-generation parts and transparent sizing. A free sizing guide for GPU memory and system RAM is at https://eracks.com/guides/private-ai-sizing/ and the full line is at https://eracks.com/products/ai-rackmount-servers/
About eRacks Open Source Systems
eRacks Open Source Systems, founded in 1999 and based in Fremont, California, specializes in open-source servers and storage. The company builds custom-configured rackmount servers, NAS (network-attached storage) systems, HPC (high-performance computing) clusters, and AI inference servers, all shipped with Linux or other open-source operating systems. More at https://eracks.com
Media Contact
Joseph Wolff eRacks Open Source Systems joe@eracks.com https://eracks.com
eRacks Open Source Systems